JOURNAL OF VIROLOGY, May 2005, p. 6291–6298
Vol. 79, No. 10
New Antiviral Pathway That Mediates Hepatitis C Virus Replicon
Interferon Sensitivity through ADAR1
Deborah R. Taylor,* Montserrat Puig, Miriam E. R. Darnell, Kathleen Mihalik, and
Stephen M. Feinstone
Laboratory of Hepatitis Viruses, Center for Biologics Evaluation and Research, Food and Drug Administration,
Bethesda, Maryland 20892
Received 25 May 2004/Accepted 1 January 2005
While many clinical hepatitis C virus (HCV) infections are resistant to alpha interferon (IFN-?) therapy,
subgenomic in vitro self-replicating HCV RNAs (HCV replicons) are characterized by marked IFN-? sensi-
tivity. IFN-? treatment of replicon-containing cells results in a rapid loss of viral RNA via translation
inhibition through double-stranded RNA-activated protein kinase (PKR) and also through a new pathway
involving RNA editing by an adenosine deaminase that acts on double-stranded RNA (ADAR1). More than 200
genes are induced by IFN-?, and yet only a few are attributed with an antiviral role. We show that inhibition
of both PKR and ADAR1 by the addition of adenovirus-associated RNA stimulates replicon expression and
reduces the amount of inosine recovered from RNA in replicon cells. Small inhibitory RNA, specific for ADAR1,
stimulated the replicon 40-fold, indicating that ADAR1 has a role in limiting replication of the viral RNA. This
is the first report of ADAR’s involvement in a potent antiviral pathway and its action to specifically eliminate
HCV RNA through adenosine to inosine editing. These results may explain successful HCV replicon clearance
by IFN-? in vitro and may provide a promising new therapeutic strategy for HCV as well as other viral
Hepatitis C virus (HCV) infects approximately 170 million
individuals worldwide and nearly 3 million in the United States
alone. Most cases of HCV infection become persistent and
may result in chronic liver disease, cirrhosis, and hepatocellu-
lar carcinoma. The current combination antiviral therapy of
pegylated alpha interferon (IFN-?) with ribavirin is effective in
approximately 50% of individuals treated, while monotherapy
with IFN-? alone is successful in less than 20% of patients (13).
IFN-? allows cells to become innately primed for defense
against eventual virus attack by inducing the transcription of
many genes, some of which are activated during virus infection.
Only a few genes have been identified and characterized as
mediators of the IFN-?-induced antiviral response, including
the Mx proteins, major histocompatibility complex proteins,
2?,5? oligoadenylate synthetase, and the double-stranded RNA
(dsRNA)-activated protein kinase (PKR) (21). PKR is acti-
vated during viral infection, which results in the phosphoryla-
tion of the ? subunit of eukaryotic translation initiation factor
2 (eIF-2?) and subsequent translational shutoff. An adenosine
deaminase that acts on dsRNA (ADAR1) is also IFN-? in-
duced and catalyzes the deamination of adenosine residues in
dsRNA (for a review, see reference 17), resulting in inosine
substitution. Inosine residues are not abundantly found in cel-
lular mRNAs, but when inosine residues are present, they are
transcribed and translated as guanosine residues, which may
lead to mutations (1, 24). An RNase that specifically degrades
inosine-containing RNA has been described and was proposed
to be part of a putative antiviral pathway (18, 19). Although
antiviral activity has not been attributed to ADAR1, hepatitis
delta virus utilizes ADAR1 editing to promote its viral life
Typically, dsRNA is found only in cells that are virus in-
fected, and both DNA and RNA viruses may present dsRNA
in the cell in the form of replicative intermediates (12).
ADAR1 contains three copies of a conserved dsRNA-binding
motif also found in PKR (25). Most viruses have developed
strategies to evade the effects of IFN-?. For instance, a single-
stranded virus-encoded RNA with partially dsRNA features,
adenovirus-associated (VA) RNAI, enhances translation and
counteracts the effects of IFN-? in adenovirus-infected cells by
inhibiting PKR (14). VA RNA has also been shown to bind
and inhibit ADAR1 (10). This is the first report implicating
RNA editing by ADAR1 in the control of viral replication and
may provide a potent strategy for an effective treatment against
HCV based on ADAR1 activation.
MATERIALS AND METHODS
Expression vectors and HCV replicon. The BB7 replicon was a gift of C. M.
Rice and K. Blight and was previously described (2). Briefly, the replicon ex-
presses HCV NS3 through NS5B nonstructural genes under the encephalomyo-
carditis virus (EMCV) internal ribosome entry site (IRES) and neomycin resis-
tance under the HCV IRES. The HCV 5? and 3? nontranslated regions are also
present. Bicistronic reporter plasmids were constructed by substitution of the
simian virus 40 promoter in pRL-SV40 (Promega) with the herpes simplex virus
type 1 alpha 27 promoter (gift of N. Martin). The mRNA expresses Renilla
luciferase (Luc) through a cap-dependent translation mechanism and firefly Luc
under the EMCV IRES. PCR of the BB7 plasmid (using primers that contained
a SalI restriction site at the 5? end and a SacI restriction site at the 3? end,
covering nucleotides 1 to 386 in pHCVrep1bBB7) (16) was used to generate an
HCV IRES cassette. The EMCV IRES was removed by digestion with SalI and
SacI, and the HCV-IRES was inserted. pcDNA3 (Invitrogen) plasmids express-
ing wild-type (WT) PKR, PKR K296R, E2, and NS5A were described previously
(20). eIF-2?-pRc/CMV (eIF-2? S51A) was a gift of O. Donze and was described
previously (4). VA RNAI-pUC119, expressed under the polymerase III intra-
genic promoter was described previously (7).
* Corresponding author. Mailing address: CBER/FDA, HFM-448,
8800 Rockville Pike, Bethesda, MD 20892. Phone: (301) 827-3660.
Fax: (301) 496-1810. E-mail: email@example.com.
Cell culture. Huh7 and Huh.BB7 cells were grown at 37°C in Dulbecco’s
modified Eagle’s medium (DMEM) with 10% fetal bovine serum, 2 mM L-
glutamine, 100 U/ml penicillin, and 100 ?g/ml streptomycin. Huh7 cells were
transfected with RNA transcribed from linearized pHCVrep1bBB7 plasmid.
Stable transfectants were selected in Geneticin (G418; Gibco) and grown in
G418 at 75 ?g/ml, which was removed during experiments. Cellular proteins
were metabolically labeled in cell culture by incubating the cells for 1 h at 37°C
in DMEM minus methionine and then supplementing the medium for 1 h with
[35S]methionine. RNA was metabolically labeled in cell culture by incubating the
cells for 1 h at 37°C in DMEM minus phosphates and then incubating the cells
for 16 h in medium supplemented with [?-32P]ATP. Cell growth was monitored
by counting the number of viable cells with trypan blue staining or with a Coulter
Counter (Beckman Coulter, Inc.). Cells were treated with recombinant human
IFN-?2A (100 IU/ml in DMEM) for 18 h at 37°C, unless stated otherwise.
Analyses of protein and RNA. Cells were washed with phosphate-buffered
saline (PBS), and the attached monolayer was incubated with trypsin EDTA for
5 minutes at 37°C. Cells were collected, concentrated in 0.1% NP-40 and 10%
PBS, and lysed by three repeated freeze-thaw cycles. Nuclear and cellular mem-
branes were removed by centrifugation, and cytoplasmic extracts were quanti-
tated for protein concentration with a colorimetric absorbance protein assay
(Bio-Rad). Expression of HCV NS3 protein in lysates from Huh7 or Huh.BB7
cells was monitored by resolving 25 ?g of protein extract in sodium dodecyl
sulfate-polyacrylamide gels and immunoblotting with polyclonal antibody to NS3
(chimp 1536 serum ). Polyclonal anti-PKR antiserum was made by inoculat-
ing New Zealand White rabbits with a keyhole limpet hemocyanin-conjugated
peptide to the spacer region between the two dsRNA-binding motifs of PKR
(peptide P1) (22). E2 expression was monitored with mouse monoclonal anti-
body A11 using cell extracts that were treated with endoglycosidase H. Actin
expression was detected using polyclonal goat antiactin antibodies (Santa Cruz
Biotechnology). Monoclonal antibodies (Cell Signaling Technology) were used
to detect phosphorylated eIF-2? and total eIF-2? proteins from 20 ?g of protein
from the cytoplasmic fraction (as determined by Bio-Rad protein assay) of cells
that were harvested as described above with the addition of phosphatase inhib-
itors (90 mM sodium fluoride, 17.5 mM sodium molybdate, 17.5 mM ?-glycer-
Cellular RNA was isolated using the Trizol method (Gibco) and quantitated
by UV absorbance at 260 nm. HCV replicon RNA from 0.5 ? 106cells was
monitored by real-time reverse transcription-PCR (RT-PCR) (Taqman; Applied
Biosystems) using primers located on the HCV 5? end and quantified on the basis
of HCV RNA standards and relative amounts of cellular glyceraldehyde-3-
phosphate dehydrogenase (GAPDH) mRNA as described previously (26).
RNase protection analysis was performed using the HybSpeed RPA kit (Am-
bion) to examine the transfection efficiency of luciferase reporters. Cellular RNA
was isolated and incubated with an antisense RNA probe to 125 nucleotides of
Renilla luciferase, synthesized with T7 RNA polymerase and [?-32P]dCTP. The
assay was performed per the manufacturer’s suggestions. RNA was quantitated
using known RNA concentrations of the sense strand of the probe, synthesized
with SP6 RNA polymerase.
Luc assays. Luciferase assays were performed with a Dual-Luciferase reporter
system (Promega) (per the manufacturer’s directions), and Luc activity was
measured with a luminometer (Turner Designs). Lysates were prepared from
Huh7 or Huh.BB7 cells 48 h after transfection with a bicistronic reporter plasmid
(2 ?g DNA/transfection) or bicistronic Luc reporter RNA (synthesized in vitro
with T7 RNA polymerase after linearization of the plasmid). Luc assay results
shown are representative of four or more experiments with transfections per-
formed in duplicate and samples tested in the Luc assay in triplicate. Similar
results were obtained in all experiments.
TLC. Radioactive monophosphates were resolved by thin-layer chromatogra-
phy (TLC) on polyethylenimine cellulose was developed in ammonium acetate
buffer as described previously (19). Briefly, RNA was digested with nuclease P1
(Roche) after Trizol isolation from Huh.BB7 cells that were grown in the pres-
ence of [?-32P]ATP. Nonradioactive AMP and IMP were used as migration
standards and visualized with UV light. Radioactivity was visualized by autora-
diography and quantitated by PhosphorImager analysis.
siRNA. Huh.BB7 cells were plated at 1 ? 106cells/ml. Huh.BB7 cells were
transfected with ADAR1 or ADAR2 small interfering RNA (siRNA) (final
concentration of 100 nM) (a gift of John Casey) as described previously (23) with
DMRIE-C transfection reagent (Invitrogen). Cells were transfected with ADAR
siRNA once or transfected twice (24 h later) and harvested 7 days after the initial
Isolation and sequencing of replicon RNA. Huh.BB7 cells were treated with
IFN-? (100 IU/ml, 18 h), and cytoplasmic RNA was extracted by disrupting cells
in 0.1% NP-40 and 0.1? (10%) PBS as described above. RT was performed using
First strand cDNA kit (Amersham) with a primer in the antisense direction of
the NS5 region (5?-CAACCGTCCTCTTCCTCCG-3?), and PCR was performed
using Expand high fidelity PCR system (Roche) with primers directed towards
the HCV IRES region (5?-GCATGCGTCGACGCCAGCCCCGGATTGGG
products were inserted into pCR II-Topo with a Topo TA cloning kit (Invitro-
gen), white clones were selected, and DNA was purified and sequenced with the
forward primer included in the kit.
IFN-? sensitivity in cell culture is replicon specific. The
mechanisms underlying IFN-? resistance by HCV have been
examined with the replicon-based system (11). Efficient repli-
cation of HCV RNA was observed in Huh7 cells that had been
stably transfected with the HCV replicon (BB7) (2, 3, 11).
While the replicons were derived from a genotype 1b strain,
which is usually the most IFN resistant, these replicons were
highly sensitive to IFN-? (2). It was even possible to com-
pletely cure the cells of the replicon with IFN-? treatment (3).
In addition, sequence adaptations that conferred robust
growth in cell culture did not reduce IFN-? sensitivity (2).
We found that expression of the HCV replicon proteins was
highly sensitive to IFN-? (Fig. 1A). Total protein synthesis in
Huh7 cells or replicon-containing cells (Huh.BB7) was also
affected, as demonstrated by the decrease in total cellular
FIG. 1. IFN-? action is replicon specific. (A) Immunoblot of HCV
NS3 protein expression in lysates (25 ?g) from parental Huh7 cells or
replicon-containing cells (Huh.BB7) treated with IFN-? (100 IU/ml)
for 24, 48, or 72 h. (B) Total cellular proteins from an equal number
of [35S]methionine metabolically labeled cells after treatment with
IFN-? (100 IU/ml) for 24, 48, or 72 h. (C) HCV RNA from 0.5 ? 106
cells was quantitated by Taqman RT-PCR using HCV standards. Total
RNA from the same cells was measured by absorbance of UV (wave-
length of 260 nm).
6292 TAYLOR ET AL.J. VIROL.
proteins synthesized after IFN-? treatment (Fig. 1B). IFN-?
treatment caused relatively little cytotoxicity in cells that con-
tained the replicon or in the parental cells (Huh7) (data not
shown). We performed Northern blot analysis to detect repli-
con RNA in IFN-?-treated cells. With increasing levels of
IFN-?, a decrease in replicon RNA was observed (data not
shown). As previously reported by Blight et al. (2), we observed
a precipitous decrease in replicon RNA after IFN-? treatment
as measured by Taqman analysis, while total RNA in the cells
was largely unaffected (Fig. 1C), demonstrating that the pri-
mary effects of IFN-? are replicon specific.
The HCV IRES is IFN-? resistant in replicon-containing
cells. Translation of HCV is initiated at the highly structured 5?
untranslated region of the viral RNA containing an IRES. To
determine how replicon-specific protein synthesis was affected
by IFN-?, we engineered a bicistronic DNA reporter that ex-
presses Renilla luciferase under a 5? cap-dependent transla-
tional mechanism and firefly Luc under a cap-independent,
IRES-directed mechanism. We examined the effects of IFN-?
on expression from both the EMCV IRES and the HCV IRES
(Fig. 2A), because both are required for replicon expression.
Both cap-dependent and EMCV IRES-dependent Luc expres-
sion was inhibited by IFN-? in both Huh7.BB7 and Huh7 cells
(Fig. 2B and C). Complete inhibition of Luc activity was not
observed, because Luc protein accumulated for 30 h prior to
the addition of IFN-?. HCV IRES-dependent Luc expression
was also inhibited by IFN-? (Fig. 2C), suggesting that there is
a global effect of IFN-? at the posttranscriptional stage in
transfected cells. This is consistent with an inhibition of trans-
lation by PKR, as PKR activation would result in a decrease in
both cap- and IRES-dependent translation.
To exclude the possibility that transcription of the Luc re-
porter was inhibited by IFN-?, we compared Luc expression
using transfected DNA to Luc expression from transfected
capped RNA synthesized in vitro from the same reporter and
observed radically different responses to IFN-? treatment.
When DNA was transfected into Huh.BB7 cells, the EMCV-
IRES-dependent expression, HCV-IRES-dependent expres-
sion, and cap-dependent expression of Luc were all inhibited
by IFN-? (Fig. 2B). The parental Huh7 cells lacking the rep-
licon behaved similarly (Fig. 2C). However, when we trans-
fected RNA encoding the Luc reporter, only HCV IRES-
dependent Luc expression was resistant to IFN-? in those cells
that contained the replicon (Fig. 2D) and was sensitive to
IFN-? in Huh7 cells (Fig. 2E). These results were repeated
with similar reporters described previously and are consistent
with those reported by Koev et al. who proposed that a PKR-
independent mechanism is responsible for IFN-? activity on
the replicon (9). Because PKR activation results in inhibition
of both cap-dependent and -independent translation, if PKR
were solely responsible for IFN-? sensitivity of the cap-depen-
dent Luc expression, then IRES-dependent Luc expression
should be affected as well. These results show that in trans-
fected cells, in the absence of the replicon, both cap-dependent
translation and cap-independent translation were inhibited
(Fig. 2C and E), which is consistent with inhibition due to
activated PKR, but this does not exclude other mechanisms of
inhibition. These results contradict results demonstrating that
the replicon is sensitive to G418 in the presence of IFN-? (data
not shown). Therefore, the HCV IRES, in the context of the
replicon, is sensitive to IFN-?. These data may suggest that the
RNA of the HCV IRES-containing reporter in the presence of
the replicon may be inhibiting or blocking the main effects of
IFN-? on IRES translation in a manner that cannot be fulfilled
by the DNA reporter or in cells that lack the replicon (Fig. 2F).
Because PKR is activated by dsRNA, which is probably present
in replicon-containing cells, translation inhibition through
PKR is expected. However, PKR did not shut down translation
FIG. 2. Cap-independent translation exhibits IFN-? resistance. (A) Bicistronic luciferase reporters expressing Renilla luciferase under control
of a cap-dependent translation mechanism and firefly luciferase under the EMCV or HCV IRES. (B and C) Luc activity expressed as fold inhibition
as a result of IFN-? treatment in cell lysates from Huh7 or Huh.BB7 cells transfected with bicistronic reporter plasmid DNA (2 ?g) containing
the EMCV (E) IRES or HCV (H) IRES constructs shown in panel A, expressed relative to the control (no IFN-?) (which was assigned a value
of 10). (D and E) IFN-? (100 IU/ml) inhibition of Luc activity in cell lysates from equal numbers of Huh.BB7 or Huh7 cells transfected with
bicistronic reporter plasmid RNA (synthesized in vitro using T7 RNA polymerase) containing the EMCV (E) or HCV (H) IRES. Luc assay results
shown are representative of four or more experiments with transfections performed in duplicate and assayed for Luc activity in triplicate.
Transfection efficiency was also monitored by RNase protection analysis (data not shown). Values are means ? standard deviations of the means
(error bars). (F) Summary of data in panels B to E.
VOL. 79, 2005 HCV REPLICON AND INTERFERON SENSITIVITY 6293
of the HCV IRES reporter in replicon cells. Taken together, it
appears that while PKR may be activated, this cannot fully
explain the IFN-? sensitivity of the replicon.
E2 and NS5A stimulate translation but did not rescue the
replicon from IFN-? sensitivity. To examine the importance of
PKR-dependent translation inhibition on the replicon, we used
PKR inhibitors to counteract the effect of IFN-? in the Luc
assay. Increasing amounts of transfected E2 stimulated both
cap-dependent and IRES-dependent translation with and
without IFN-? (Fig. 3A). Stimulation of translation corre-
sponded with increasing levels of E2 from cells cotransfected
with the Luc reporter. Equal amounts of Luc reporter were
transfected in these cells, as demonstrated by the RNase pro-
tection assay (Fig. 3A, RPA). Both E2 and NS5A have been
shown to inhibit PKR (17, 18), and expression of each of these
HCV genes enhanced Luc expression in the presence and
absence of exogenous IFN-? (Fig. 3B). We used a high level of
IFN-? to show the enhanced effect of E2 and NS5A. No dif-
ferences in reporter expression were seen in Huh7 cells trans-
fected with vector DNA. Interestingly, the combination of both
E2 and NS5A enhanced Luc expression in Huh.BB7 cells in an
additive manner (Fig. 3B), suggesting that E2 and NS5A coun-
teract the effects of IFN-? in different and complementary
ways. Both E2 and NS5A were capable of rescuing the repli-
cation of HCV replicon RNA during IFN-? treatment. Com-
plete rescue was not observed, as the amount of HCV RNA
declined after 24 h of IFN-? treatment (Fig. 3C).
PKR is not induced but is activated by dsRNA present in
replicon-containing cells. PKR was fully induced by IFN-?
treatment in replicon-expressing cells (Fig. 3D), but PKR was
not endogenously induced in the absence of IFN-? as indicated
by the low level of PKR expression seen in untreated cells and
the induction of PKR expression after IFN treatment. Endog-
enous IFN-? or IFN-? was, therefore, not being secreted by
untreated Huh.BB7 cells. PKR activation was monitored by
observing phosphorylation levels of PKR substrate eIF-2?
(Fig. 3E) compared to the total amount of eIF-2? (Fig. 3E). In
Huh7 cells, eIF-2? was phosphorylated with the addition of
IFN-? and dsRNA, but very little eIF-2? was phosphorylated
in the absence of treatment (Fig. 3E, lanes 1 and 2). This is
consistent with the dsRNA-dependent activation of PKR. In
Huh.BB7 cells, however, a high level of eIF-2? phosphoryla-
tion was observed in the absence of IFN-?, and no additional
phosphorylation was observed with the addition of dsRNA or
FIG. 3. Inhibition of PKR stimulates translation but does not result in complete rescue of the replicon from the effects of IFN-?. (A) Luciferase
activity measured in cell lysates from Huh.BB7 cells transfected with HCV IRES bicistronic reporter plasmid (2 ?g) and E2 and/or pcDNA3
plasmids (10 ?g total DNA/6-cm dish). Cells were untreated [(?) IFN] or treated with 1,500 IU/ml IFN-? [(?) IFN] for 18 h. The expression levels
of transfected E2 are shown below the bars. Immunoblot of E2 from transfected cells. RNA was isolated from transfected cells, and Luc
expression/transfection efficiency was monitored by RNase protection analysis (RPA) as described in Materials and Methods. (B) Luciferase
activity in cells cotransfected with HCV IRES bicistronic reporter plasmid and E2 or NS5A (6 ?g) with pcDNA3 (6 ?g) or with both E2 and NS5A
(E?N; 6 ?g each). Luciferase activity is expressed as fold increase. Values are means ? standard deviations of the means (error bars).
(C) Huh.BB7 cells transfected in triplicate with E2, NS5A, or pcDNA3 (vector). Cells were untreated (no IFN-?) or treated with 100 IU/ml IFN-?
at 8 h posttransfection (72 h), 32 h posttransfection (48 h), or 56 h posttransfection (24 h) and were harvested at 80 h posttransfection. Replicon
RNA was measured by Taqman analysis of HCV RNA relative to the level of GAPDH mRNA in 0.5 ? 106cells (16). Values are means ? standard
deviations of the means (error bars). (D) Cells transfected with pcDNA3 vector (V) or E2 pcDNA3 (E2), were treated with 1,000 IU/ml IFN-?
for 18 h (?) and analyzed for PKR or actin expression by protein immunoblotting with polyclonal anti-PKR antiserum or antiactin antibodies.
(E) Immunoblot detection of phosphorylated eIF-2? (phospho-eIF-2?) and total eIF-2? (20 ?g total protein/lane) from cells transfected with E2
or NS5A or E2 plus NS5A and treated with 1,000 IU/ml IFN-? and poly(I · C) (20 ?g/ml) for 18 h (?).
6294 TAYLOR ET AL.J. VIROL.
IFN-? (Fig. 3E, lanes 3 and 4). This indicates that although
PKR was at a low basal level of expression in untreated Hu-
h.BB7 cells (Fig. 3D), it was already activated in replicon-
containing cells (Fig. 3E, lanes 3 and 4). Because no additional
eIF-2? phosphorylation was observed after IFN-? and dsRNA
treatment (Fig. 3E, compare lanes 3 to 4), eIF-2? phosphory-
lation may have already saturated. Alternatively, the assay it-
self may be saturated, allowing for no additional detection of
eIF-2? phosphorylation. To test this possibility and to evaluate
the importance of PKR activation, we transfected the cells with
the known PKR inhibitors HCV E2 (20) or NS5A (6) or both
E2 and NS5A. All samples containing transfected E2 or NS5A
demonstrated a decrease in the level of phosphorylated eIF-2?
(Fig. 3E, compare lanes 5 to 10 to lanes 3 and 4), confirming
that these two HCV genes inhibit PKR. Still, there was no
additional phosphorylation of eIF-2? in the presence of IFN-?
(Fig. 3E), even though translation decreased with IFN-? (Fig.
3A) and PKR protein expression increased in these cells (Fig.
3D), confirming that in these samples the assay was not satu-
rated. These results also suggest that IFN-? treatment of rep-
licon-containing cells does not lead to additional activation of
PKR. This supports evidence demonstrated here and by others
(9) suggesting that not only PKR but other pathways modulate
IFN-? sensitivity of the replicon. It is consistent with the re-
sults showing that the PKR inhibitors can stimulate the repli-
con but cannot overcome all the negative effects of IFN-?
treatment (Fig. 3B, C, and E).
Because PKR was not fully induced (Fig. 3D) yet was al-
ready activated in the cells expressing the replicon (Fig. 3E),
dsRNA was most likely present, presumably from RNA repli-
cation of both positive- and negative-strand RNA. Taken to-
gether, these data are consistent with a stimulation of replicon
expression by E2 and NS5A due to inhibition of PKR, which is
activated by the replicon. Because IFN-? treatment was not
required for activation of PKR (or its downstream effects) in
replicon-containing cells (Fig. 3E) and yet, IFN-? is a potent
antagonist of the replicon, a second IFN-?-induced antiviral
pathway may be involved. This is also consistent with the re-
sults obtained in Fig. 2, where HCV IRES translation was
insensitive to IFN-?. As shown previously by Koev et al. (9),
PKR activation should result in the inhibition of both IRES-
mediated and cap-mediated translation. Therefore, inhibition
of the replicon by IFN may proceed through a secondary
mechanism to PKR activation. These data may suggest that the
low level of PKR expression was sufficiently activated to enable
phosphorylation of maximum levels of eIF-2?.
Activation of PKR probably occurred as a result of dsRNA
present in replicon-containing cells but not in Huh7 cells alone.
This is consistent with other findings showing that IFN-? reg-
ulatory factor 3 (IRF-3), which is important for the initial
induction of IFN-? in response to dsRNA or virus infection of
cells, was inhibited in the presence of the replicon (5). The
activation of IRF-3 usually leads to the induction of IFN-? and
subsequently of IFN-?/?-induced genes, such as PKR. We
found that while dsRNA was present in replicon-containing
cells, PKR expression was not induced, suggesting that the
dsRNA induction of IFN-? was blocked, which is consistent
with an inhibition of IRF-3.
VA RNAIconfers IFN-? resistance to the replicon. To ex-
amine the importance of eIF-2? phosphorylation in cellular
and viral translation, we tested IFN-? sensitivity of a dual-Luc
reporter. We used WT PKR and well-characterized PKR in-
hibitors to measure stimulation of translation, which was mea-
sured as an output of Luc expression. WT PKR, a catalytically
inactive dominant-negative mutant PKR (PKR K296R), an
eIF-2? phosphorylation site mutant (eIF-2? S51A) that is non-
phosphorylatable, and VA RNAIwere cotransfected with the
HCV IRES Luc reporter into Huh.BB7 cells. Overexpression
of PKR (WT) inhibited expression of the Luc reporter, as
expected (Fig. 4A), consistent with the behavior of transfected,
active PKR (22). Both PKR K296R and eIF-2? S51A stimu-
lated Luc expression, even in the presence of IFN (Fig. 4A),
suggesting that eIF-2? phosphorylation was involved in the
inhibition of both IRES-dependent and cap-dependent trans-
lation by IFN-?. Cotransfection with a DNA plasmid encoding
VA RNAIresulted in 25-fold stimulation of IRES-directed
Luc expression in the absence of IFN and nearly 20-fold stim-
ulation in the presence of IFN (Fig. 4A). Cap-dependent trans-
lation was also stimulated by VA and to a higher degree than
demonstrated by any of the other PKR inhibitors. While VA
RNA is a potent inhibitor of PKR, this finding may suggest that
VA RNAIcan stimulate translation in replicon-containing
cells by conferring IFN resistance through a pathway in addi-
tion to PKR. To examine the effects of these inhibitors on
replicon RNA expression, we measured HCV replicon RNA in
cells that were treated with IFN-? and transfected with the
PKR inhibitors. Replicon expression was stimulated by all of
the inhibitors but was stimulated most strongly in the presence
of VA RNAIand eIF-2? S51A (Fig. 4B). When these cells
were treated with IFN-?, all inhibitors stimulated replicon
FIG. 4. VA RNAIrescues the replicon from the effects of IFN-?.
(A) Luciferase activity in Huh.BB7 cells cotransfected with HCV IRES
bicistronic reporter and pcDNA3, wild-type PKR, PKR K296R, eIF-2?
S51A, or VA RNAI. Cells were treated with 500 IU/ml IFN-? for 18 h.
Luc activity is relative to the pcDNA3 control value, which was as-
signed a value of 1. (B) Taqman quantitation of HCV RNA from
Huh.BB7 cells transfected with pcDNA3 (vector) or PKR inhibitors.
Cells were untreated or treated with IFN-? at 500 IU/ml for 18 h (?).
The histogram shows the fold increase in RNA detected from an equal
number of inhibitor-transfected cells relative to vector-transfected
cells, which was assigned a value of 1, relative to the amount of
GAPDH mRNA. Values are means ? standard deviations of the
means (error bars). Below the histogram are the actual RNA copy
numbers (106) per reaction based on Taqman HCV RNA standards.
VOL. 79, 2005HCV REPLICON AND INTERFERON SENSITIVITY6295
expression over vector alone or the catalytically inactive PKR
K296R mutant. VA and NS5A, however, showed very efficient
rescue of replicon RNA expression (Fig. 4B). Because the
PKR inhibitors eIF-2? S51A and PKR K296R stimulated the
replicon weakly (Fig. 4B), it is evident that PKR is involved in
the limitation of the replicon in these cells. However, the
robust stimulation of translation (Fig. 4A) and replicon RNA
by VA RNAI(Fig. 4B) may suggest that a second potent
antiviral pathway may be involved in the IFN-?-induced inhi-
bition of the HCV replicon. Because VA RNAIbinds and
inhibits ADAR1 in vitro (10), we looked for evidence of
ADAR1 activity in replicon-containing cells.
IFN-? results in A-to-I mutations in replicon RNA. To ex-
amine evidence for editing in IFN-?-treated replicon cells, we
sequenced RT-PCR products from cytoplasmic fractions of
IFN-?-treated and untreated Huh.BB7 cells. If adenosine-to-
inosine editing events occurred, then adenosine residues will
read as guanosine residues. None of the clones from untreated
cells, sequenced from three distinct regions of the replicon
(data not shown), diverged from the WT replicon sequence.
However, one clone obtained from the IFN-?-treated cells
contained mutations in adenosine residues resulting in
guanosine (Fig. 5), suggesting that the replicon RNA was di-
rectly edited. Because we had difficulty in obtaining PCR prod-
ucts that contained mutations, this suggests that once se-
quences are edited, they may not be replicated or they may be
VA RNAIimpairs A-to-I editing in replicon cells. In addition
to its inhibition of PKR, VA RNAIis also known to inhibit
ADAR1 (10). ADAR1 acts specifically on dsRNA, is IFN-?
inducible, and causes the conversion of adenosine in dsRNA to
inosine by deamination. To test the hypothesis that RNA ed-
iting occurs in replicon-containing cells, we examined the ef-
fects of IFN-? treatment on conversion of radiolabeled AMP
to IMP. We isolated RNA that was metabolically labeled with
[?-32P]ATP from the cytoplasmic fraction of IFN-?-treated
Huh.BB7 cells. Because VA RNAIhas been shown to specif-
ically inhibit ADAR1 (10) and can also rescue IFN-?-inhibited
replicon expression very efficiently (Fig. 4B), we wanted to
know whether VA was abrogating IFN-? sensitivity by inhib-
iting ADAR in replicon-containing cells. We measured radio-
labeled IMP production in cytoplasmic RNA from Huh.BB7
cells transfected with VA RNAIby PhosphorImager quantita-
tion (Fig. 6A). Radiolabeled IMP was efficiently produced
(16% of total adenosine products) in IFN-treated replicon
cells. When we transfected the plasmid encoding VA RNAI,
the generation of radiolabeled IMP was inhibited to less than
1%, consistent with a role for VA in the inhibition of A-to-I
editing by ADAR1 in IFN-?-treated replicon-containing cells.
Previously, ADAR has been shown to convert approximately
40% of adenosines to inosines in dsRNA (19).
In VA RNA-transfected cells, the replicon was stimulated in
the presence and absence of IFN-? (Fig. 6B), and yet ADAR1
was upregulated only when the cells were treated with IFN-?
(Fig. 6C). This is consistent with previous observations
FIG. 5. IFN-?-treated Huh.BB7 cells yield edited replicon RNA.
Four regions within the HCV IRES of BB7 replicon that contain
mutations are shown. The wild-type BB7 sequences are shown. The
nucleotides in the wild-type BB7 sequence that were mutated are
shown in bold type and underlined, and the nucleotides obtained after
IFN-? treatment are shown below the sequence.
FIG. 6. A-to-I RNA editing in replicon-containing cells. (A) Thin-
layer chromatography of nuclease P1-digested RNA from Huh.BB7
cells grown in the presence of [?-32P]ATP with (?) and without (?)
100 IU/ml IFN-? in the absence (?) and presence (?) of transfected
VA RNAIplasmid. Monophosphates were resolved on one TLC plate,
although the rightmost lane was exposed longer. Radioactivity was
quantitated by PhosphorImager analysis and is shown below the gel as
a percentage of the total counts (IMP, ATP, AMP, and origin).
(B) Replicon-containing cells were transfected with VA RNA and
treated with IFN-? (?). HCV replicon RNA was monitored by Taq-
man analysis and was quantitated relative to the amount of GAPDH
mRNA (16). Values are means ? standard deviations of the means
(error bars). Transfections were performed in duplicate with one dish
of cells used for Taqman analysis and the other used for immunoblot
analysis (C) where cytoplasmic extracts (20 ?g) were used to monitor
ADAR1 and actin expression.
6296 TAYLOR ET AL.J. VIROL.
whereby replicon expression and Luc expression were stimu-
lated by VA RNA in the absence of IFN-? in replicon cells
(Fig. 4) and suggests that ADAR1 may already be active (as
with PKR) due to dsRNA present in replicon-containing cells.
siRNA knockdown of ADAR1 stimulates replicon expres-
sion. To confirm that the RNA editing occurring in replicon-
containing cells was attributable to ADAR1, we used an RNA
interference assay utilizing siRNA specifically directed to
knock down the expression of ADAR1 (8). We used a siRNA
that has been shown upon transfection to markedly decrease
ADAR1 expression specifically (8). Cells were transfected
once, and after 24 h, some cells were transfected for a second
time. HCV replicon RNA increased with transfection of
ADAR1 siRNA (Fig. 7A). At day 7 posttransfection, the HCV
replicon RNA increased by 41-fold for one transfection and
fivefold for two transfections of ADAR1 siRNA (Fig. 7A).
Two transfections of ADAR1 siRNA were slightly toxic to the
cells (data not shown), which may be caused by the loss of
activity of the siRNA. The results observed directly correlate
with the expression level of the p150, cytoplasmic, IFN-?-
induced form of ADAR1 in cells (Fig. 7B). siRNA directed to
knockdown ADAR2 (Fig. 7A) showed no stimulation of HCV
replicon RNA, suggesting that ADAR1 and not ADAR2 is
responsible for instability of HCV replicon RNA.
The stimulation of the replicon by ADAR1 knockdown was
seen in the absence of IFN-? (Fig. 7A and B). We, therefore,
tested the effects of siRNA directed to ADAR1 in the presence
of IFN-?. In the absence of siRNA transfection, IFN-? treat-
ment resulted in a decrease in replicon RNA (Fig. 7C) corre-
sponding with an increase in ADAR1 expression (Fig. 7D).
Transfection with siRNA directed to ADAR1 resulted in loss
of replicon RNA during IFN-? treatment as well (Fig. 7C).
However, compared to no-siRNA controls, the samples with
siRNA yielded more replicon RNA (Fig. 7C), supporting the
conclusion that IFN-? sensitivity of the replicon is mediated
through ADAR1. The ADAR1 immunoblot (Fig. 7D) demon-
strates that in the presence of IFN-?, siRNA does not com-
pletely knock down ADAR1 expression, consistent with the
observed replicon sensitivity in these samples.
Our findings point to a new IFN-?-induced antiviral pathway
important to the modulation of HCV replicon replication in
cell culture. This pathway involves dsRNA-specific editing of
adenosine residues by ADAR1 in HCV replicon-containing
cells, which then leads to the loss of HCV replicon RNA. Viral
RNA clearance may be attributable to one or several factors.
(i) The viral RNA may be edited and degraded, possibly by an
inosine-specific RNase (18, 19). (ii) Inefficient replication and
genome instability may result from mutation of important viral
sequences. (iii) A cellular message that is required for viral
replication may be edited and inactivated. Taken together, our
results explain why IFN-? action is specific for HCV replicon
RNA and not cellular RNA and how VA RNAIcan rescue
replicon expression so effectively. We are currently testing the
HCV IRES RNA for its ability to bind and inhibit ADAR1
function. Why the addition of HCV IRES-containing RNA
confers IFN-? resistance to IRES-dependent translation and
not cap-dependent translation is not easily answered, although
it may indicate that the HCV IRES itself is protected from an
IFN-?-induced mechanism (23). These results confirm the
findings that IFN-? action on the replicon is a result of both
PKR and ADAR1 activation.
At this time we have not identified a specific editing site in
HCV RNA, but random hyperedited viral RNA may be the
result of ADAR1 action on the replicon. Because of the high
level of radioactive inosine obtained, the HCV replicon may be
hyperedited as a result of dsRNA produced by the replicase.
At this time it has been difficult to obtain RT-PCR products
containing mutations in the replicon that represent edited ad-
enosines, and RT-PCR products from IFN-?-treated replicon
cells have not yielded an abundance of clones. We believe that
this may be due to the inability of the mutated replicon to
replicate, the inability of PCR primers to recognize mutated
sequences, and/or the HCV replicon RNA may be quickly
degraded after IFN-? treatment.
To what extent ADAR is involved in IFN-? sensitivity in
patients is also not understood at this time. The presence and
genetic variability of viral genes that interact with ADAR1 may
play a role in the outcome of IFN-? therapy. Finally, the
discovery that RNA editing negatively affects HCV RNA rep-
FIG. 7. siRNA knockdown of ADAR1 stimulates replicon expres-
sion. (A) Huh.BB7 cells were plated at 1 ? 106and transfected once
or twice with siRNA directed towards ADAR1 or with siRNA directed
against ADAR2. The histogram shows Taqman analysis of HCV RNA
from equal numbers of cells. The number of transfections is shown
below the bars in the histogram. (?) Twenty-five micrograms of cyto-
plasmic lysates shown in panel A. siRNA experiments were repeated
three times with similar results. (C) Huh.BB7 cells were transfected
with ADAR1-specific siRNA and treated with IFN-? (10 or 1,000
IU/ml) at 4 days posttransfection. Cells were harvested after 18 h of
IFN-? treatment. RNA was isolated and monitored by Taqman anal-
ysis (16). (D) Cytoplasmic extracts were analyzed for protein expres-
sion as described in Materials and Methods. The migration positions of
molecular mass markers are shown to the left of the blots.
VOL. 79, 2005 HCV REPLICON AND INTERFERON SENSITIVITY6297
lication may lead to new approaches in the development of
therapies that target this new IFN-?-induced antiviral pathway
important for the clearance of HCV and potentially other
We thank John Casey for ADAR siRNAs and ADAR1 antibody,
Charles Rice and Keril Blight for replicon cell lines and DNA, Gen-
nadiy Koev and Michael Lai for Luc plasmids and Natalia Martin for
EMCV Luc plasmid, Olivier Donze for the mutant eIF-2? plasmid,
and Nahum Sonenberg, Barry Falgout and Geetha Jayan for com-
ments on the manuscript. IFN-? was supplied by the NIH Research
and Reference Reagent Program of NIAID.
1. Bass, B. L. 1997. RNA editing and hypermutation by adenosine deamina-
tion. Trends Biochem. Sci. 22:157–162.
2. Blight, K. J., A. A. Kolykhalov, and C. M. Rice. 2000. Efficient initiation of
HCV RNA replication in cell culture. Science 290:1972–1974.
3. Blight, K. J., J. A. McKeating, and C. M. Rice. 2002. Highly permissive cell
lines for subgenomic and genomic hepatitis C virus RNA replication. J. Vi-
4. Donze, O., R. Jagus, A. E. Koromilas, J. W. Hershey, and N. Sonenberg.
1995. Abrogation of translation initiation factor eIF-2 phosphorylation
causes malignant transformation of NIH 3T3 cells. EMBO J. 14:3828–3834.
5. Foy, E., K. Li, C. Wang, R. Sumpter, Jr., M. Ikeda, S. M. Lemon, and M. J.
Gale. 2003. Regulation of interferon regulatory factor-3 by the hepatitis C
virus serine protease. Science 300:1145–1148.
6. Gale, M. J., M. J. Korth, N. M. Tang, S.-L. Tan, D. A. Hopkins, T. E. Dever,
S. J. Polyak, D. R. Gretch, and M. G. Katze. 1997. Evidence that hepatitis C
virus resistance to interferon is mediated through repression of the PKR
protein kinase by the nonstructural 5A protein. Virology 230:217–227.
7. Gunnery, S., and M. B. Mathews. 1995. Functional mRNA can be generated
by RNA polymerase III. Mol. Cell. Biol. 15:3597–3607.
8. Jayan, G. C., and J. L. Casey. 2002. Inhibition of hepatitis delta virus RNA
editing by short inhibitory RNA-mediated knockdown of ADAR1 but not
ADAR2 expression. J. Virol. 76:12399–12404.
9. Koev, G., R. F. Duncan, and M. M. C. Lai. 2002. Hepatitis C virus IRES-
dependent translation is insensitive to an eIF2?-independent mechanism of
inhibition by interferon in hepatocyte cell lines. Virology 297:195–202.
10. Lei, M., Y. Liu, and C. E. Samuel. 1998. Adenovirus VAI RNA antagonizes
the RNA-editing activity of the ADAR adenosine deaminase. Virology 245:
11. Lohmann, V., F. Korner, J.-O. Koch, U. Herian, L. Theilmann, and R.
Bartenschlager. 1999. Replication of subgenomic hepatitis C virus RNAs in
a hepatoma cell line. Science 285:110–113.
12. Maran, A., and M. B. Mathews. 1988. Characterization of the dsRNA im-
plicated in the inhibition of protein synthesis in cells infected with a mutant
adenovirus defective for VA RNAI. Virology 164:106–113.
13. McHutchison, J. G., S. C. Gordon, and E. R. Schiff. 1998. Interferon alfa-2b
alone or in combination with ribavirin as initial treatment for chronic hep-
atitis C. N. Engl. J. Med. 339:1485–1492.
14. O’Malley, R. P., T. M. Mariano, J. Siekierka, and M. B. Mathews. 1986. A
mechanism for the control of protein synthesis by adenovirus VA RNAI. Cell
15. Polson, A. G., B. L. Bass, and J. L. Casey. 1996. RNA editing of hepatitis
delta virus antigenome by dsRNA-adenosine deaminase. Nature 380:454–
16. Puig, M., M. E. Major, K. Mihalik, and S. M. Feinstone. 2004. Immunization
of chimpanzees with an envelope protein-based vaccine enhances specific
humoral and cellular immune responses that delay hepatitis C virus infec-
tion. Vaccine 22:991–1000.
17. Saunders, L. R., and G. N. Barber. 2003. The dsRNA binding protein family:
critical roles, diverse cellular functions. FASEB J. 17:961–983.
18. Scadden, A. D. J., and C. W. J. Smith. 1997. A ribonuclease specific for
inosine-containing RNA: a potential role in antiviral defense? EMBO J.
19. Scadden, A. D. J., and C. W. J. Smith. 2001. Specific cleavage of hyper-edited
dsRNAs. EMBO J. 20:4243–4252.
20. Taylor, D. R., S. T. Shi, P. R. Romano, G. N. Barber, and M. M. C. Lai. 1999.
Inhibition of the interferon-inducible protein kinase PKR by HCV E2 pro-
tein. Science 285:107–110.
21. Taylor, D. R. 2000. Hepatitis C virus: evasion of the interferon-induced
antiviral response. J. Mol. Med. 78:182–190.
22. Taylor, D. R., B. Tian, P. R. Romano, A. G. Hinnebusch, M. M. C. Lai, and
M. B. Mathews. 2001. HCV E2 envelope protein does not inhibit PKR by
simple competition with autophosphorylation sites in the RNA-binding do-
main. J. Virol. 75:1265–1273.
23. Vyas, J., A. Elia, and M. J. Clemens. 2003. Inhibition of the protein kinase
PKR by the internal ribosome entry site of hepatitis C virus genomic RNA.
24. Wang, Q., J. Khillan, P. Gadue, and K. Nishikura. 2000. Requirement of the
RNA editing deaminase ADAR1 gene for embryonic erythropoiesis. Science
25. Wong, S. K., and D. W. Lazinski. 2002. Replicating hepatitis delta virus RNA
is edited in the nucleus by the small form of ADAR1. Proc. Natl. Acad. Sci.
26. Wong, S. K., S. Sato, and D. W. Lazinski. 2003. Elevated activity of the large
form of ADAR1 in vivo: very efficient RNA editing occurs in the cytoplasm.
6298 TAYLOR ET AL. J. VIROL.